# Faisal Md

Faisal Md shares practical software-engineering lessons on using AI coding tools with human oversight, choosing system architectures that fit a team's size, and reducing security risks in backend systems.

- Profile: https://b2binfluencer.marketing/creators/faisal-md/
- LinkedIn: https://www.linkedin.com/in/se-faisal
- LinkedIn followers: 622
- Topics: artificial-intelligence, leadership, cybersecurity
- Based on 10 dated posts, latest 2026-10-06

## Summary

Faisal Md is most useful for practical technical content on AI coding oversight, right-sized architecture, and backend security trade-offs.

- Best for: A B2B software consultancy commissioning a technical collaboration series on AI coding and system design.
- Not for: An AI productivity vendor seeking reach-led endorsement, paid testimonial, or guaranteed purchase recommendation.
- Check first: Confirm recent audience demographics, per-post impressions, engagement, and whether they accept partnership work.

### Takeaways

- An AI-built e-commerce shop can look finished while cart and data handling are broken.
- Microservices can help scaling and isolation but burden a six-person team with coordination and debugging.
- Plain-text backend logs can expose sensitive data, and phishing can steal credentials despite expensive defences.

## Main ideas

- **AI can write the code, but someone still has to own the architecture**: The argument is that AI coding tools are genuinely useful for repetitive, low-level work, but should be managed as supervised juniors rather than autonomous builders. Directing the work module by module keeps the build organised, but the output needs line-by-line review: a finished-looking interface says nothing about whether the underlying business logic exists. Teams can outsource typing, not critical thinking; the senior engineer remains responsible for the architecture the tool misses.
- **Match the architecture to the team that has to run it, not to what the pattern promises**: Faisal Md treats service separation as a sound idea that becomes a liability when adopted ahead of a team's size. The benefits are clear — independent deploys, targeted scaling, and one service failing without taking checkout down — as are the coordination costs: simple changes spanning several services, ordered deploys, faults that surface only across service boundaries, and feature work giving way to plumbing. The conclusion is that the architecture was designed for a much larger company than the one operating it, not that the pattern itself was wrong.
- **Data you never collect is risk you never carry**: The recurring position is that convenience-driven logging quietly creates a second, unhardened store of sensitive information, and the fix is data minimisation rather than better protection of the logs. If a system records everything, it will eventually record credentials, payment details, and private conversations; attackers can read the log files instead of attacking the protected database. The recommended practice is to stop collecting what is not needed and automatically filter sensitive values before anything is written to disk.
- **Expensive defences fail at the human layer**: Faisal Md challenges the assumption that major breaches come from sophisticated zero-day exploits in enterprise firewalls, locating the weak point instead in ordinary user behaviour. The attack does not drop obvious malware; it captures valid credentials and an active session, after which security monitoring sees nothing unusual because the access is legitimate. The conclusion: software defences have limited value while the human layer stays unguarded, making user-level awareness and verification a priority.
- **Raises trade-offs rather than recommending tools**: Across the posts, the writer presents the case for a practice fairly, then shows what it costs, leaving the judgment to the reader rather than endorsing a product or method. The topics are framed as balances — AI speed against code quality, architectural flexibility against operational overhead, fast incident response against data exposure, software defence against human judgment — and the writing stays with the scenario and the decision rather than tool promotion. Three of the four posts close with a direct question inviting practitioners to share their experience, keeping the pieces open to discussion rather than presenting advice.

## Work to read

- [Vibe coding an e-commerce app with Antigravity IDE](https://www.linkedin.com/posts/se-faisal_vibecoding-antigravityide-softwareengineering-activity-7417497084062486528-nD50)
- [Why a six-person team should not split into microservices too early](https://www.linkedin.com/posts/se-faisal_softwareengineering-systemdesign-techleadership-activity-7511101017137029121-Hj32)
- [The hidden risk in logging every user action](https://www.linkedin.com/posts/se-faisal_backenddevelopment-cybersecurity-dataprivacy-activity-7504835955644739584-EoQJ)
- [Most breaches start with a normal-looking phishing link, not a zero-day](https://www.linkedin.com/posts/se-faisal_cybersecurity-informationsecurity-techleadership-activity-7513144691664564224-293_)

## Brand relationships

- **Antigravity.Google** (mention)

## Partnerships

Four LinkedIn posts cover software engineering: using AI coding tools under human architectural review, choosing system architecture to fit team size, auditing what backend logs store, and how credential phishing defeats expensive security tooling. None describes a sponsorship offer, paid partnership, media kit, rate card, format list, brand collaboration example, or official inquiry route for partnership work. Buyers can approach Faisal Md through the public LinkedIn profile; ask whether they accept partnership work and negotiate terms, formats, and deliverables from scratch.

## Alternatives

- Md. Jony Hassain: Md. Jony Hassain is the stronger choice when the campaign calls for a hands-on offensive-security demonstration rather than a coding-tool review. In the posts, Hassain walks through a business-logic flaw in a test platform: a balance feature allowed a negative value, a manipulated negative balance, and checkout logic still completed orders. That gives the consultancy a concrete vulnerability narrative—front-end limits versus missing server-side validation—for a security-focused episode or workshop on why feature work and security testing must be scheduled together. A second post supplies a cost case: an e-commerce site launches features without security testing and faces a $2.5M breach, lost trust, and legal penalties.
- Kareem Adebisi: Kareem Adebisi is the better fit when the buyer's task shifts toward security awareness for a general business audience. The posts tell accessible breach stories—reused passwords turning one breach into account takeover, and a lookalike partner email leading to a $50k wire transfer sent overseas—and each ends with a practical habit: use a password manager or verify through a second channel. The writing carries a clear organisational voice and recurring newsletter call to action, which can suit a consultancy selling security training or managed awareness.
- Robert Horne: Robert Horne is the stronger pick when the campaign calls for strategic security architecture rather than hands-on coding oversight. The posts argue that layered defence is not a bigger moat: firewalls keep most attackers out, but once inside, lateral movement and credential theft do the damage. They also argue that Kubernetes is rarely the target itself, but the fastest route to identities and cloud resources. Together, the posts give a consultancy a zero-trust and assume-breach narrative for a series on architecture decisions at scale, including runtime visibility and east-west traffic.

---
AI-assisted summary of public posts. Audience demographics are not available. Engagement appears only where we checked recent posts. Method: https://b2binfluencer.marketing/methodology/